Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
SlideShare a Scribd company logo

Trying Not to Filter: Internet Filtering Technologies in Libraries

•Download as PPT, PDF•
•

A presentation discussing the filtering challenge presented at the San Jose Public Library from its City Council (which the library eventually won to keep the libraries filter-free). The results of an intensive year-long study and filtering testing process are included in the presentation, as well as some information on how filters work and what to look for if you have filters in your institution.

1 of 44
Trying Not to Filter: Internet Filtering Technologies Update Sarah Houghton-Jan Digital Futures Manager for San Jose Public Library author of LibrarianInBlack.net This presentation will be at LibrarianInBlack.net
San Jose Public Library's Filtering Challenge 1 City Council member + the Values Advocacy Council (& their $)‏ Proposed a new Internet Use Policy that would require filtering on all public computers Purpose stated: “to block websites that contain child pornography or material that is obscene”
What is child pornography? What is harmful and obscene? There is no legal definition of pornography
Federal Code & Case Law Applicable Court Cases Reno v. ACLU (1998), Miller v. California (1973), Smith v. United States (1977), Pope v. Illinois (1987)‏ Applicable Federal Law 18 U.S.C. Section 1470; 47 U.S.C. section 223(d) –Communications Decency Act of 1996, as amended by the PROTECT Act of 2003; 47 U.S.C. section 231 –Child Online Protection Act of 1998
California State Code State Penal Code 311 defines obscene matter as “matter, taken as a whole, that to the average person, applying contemporary statewide standards, appeals to the prurient interest, that, taken as a whole, depicts or describes sexual conduct in a patently offensive way, and that, taken as a whole, lacks serious literary, artistic, political, or scientific value.”
California State Code State Penal Code 313 defines harmful matter as “matter, taken as a whole, which to the average person, applying contemporary statewide standards, appeals to the prurient interest, and is matter which, taken as a whole, depicts or describes in a patently offensive way sexual conduct and which, taken as a whole, lacks serious literary, artistic, political, or scientific value for minors.”
“ No filter, however, actually limits its categories to obscene material and child pornography because the current definition of obscenity doesn’t work on the internet.” ~ Lori Bowen Ayre, “Filtering and Filter Software”
The Library’s Plan Work with the community to determine their position Assess cost and impact of filter implementation Internal policy review Review of other libraries’ practices Testing filters to see if they have improved
Working with the Community Academic Senate from SJSU wrote a formal resolution against filtering Youth Commission – against Library Commission – against Public Feedback – against (about 3/4)‏
Assessing Cost & Staffing Impact The money we'd get from government paled in comparison to the cost of implementation & maintenance. ERATE isn't worth it!
Internal Policy & Other Libraries Existing computer use policy Talking with local libraries & libraries mentioned specifically by the city council member
Testing Filters
How Filters Work with URL List of trigger webpages Searches for trigger words in a popular search engine to form list (the black list!)‏ Some companies spot-check, some don't
How Filters Work with Content List of trigger words Manually generated lists of words, based on thesauruses, combined with other factors (banner ads, numbers of images and/or links, etc.). Automated search process looks for a formula of the above then lists pages. Small # of filters filter based on file type
Local Control Some filters let you form a white list of “allowed but normally blocked” content (URLs or keywords)‏ Some filters let you add to the blacklist Some filters let you temporarily unblock something for a customer.
Major Filters on the Market 602LAN SUITE Content Filter 8e6 Barracuda Bess SmartFilter ContentProtect Professional Suite CyberPatrol CyberSentinel CyberSetting CyberSitter
DP Inspector Dan’s Guardian DynaComm i:filter eSoft Web ThreatPak FastTracker FilterGate FilterPak FortiGate Series Image-Filter
INternet Filter IF-2K iPrism Netmop MaxProtect NetNanny NetSentron NetSpective Webfilter NetSweeper Network Guardian
Poesia SafeEyes Squidgard SurfControl Web Filter SurfPass WebSense Website-Echo WiseChoice
What is being filtered, exactly? Filtering software companies do not tell their customers, in detail, the types of things they block in each category. Customers cannot obtain a complete list of words or URLs that are being blocked
Choosing Filters to Test Working with IT Test computers on a test network Reviewing filter product information Reviewing filters being used in libraries Reviewing filter research and literature
Filters We Tested WebSense Enterprise CyberPatrol Filtergate Barracuda
What We Tested 135 questions/scenarios general keyword searches direct URL access image searches email text and photo attachments RSS feed content access searches in the online library catalog and databases
How We Tested 2 librarians assigned to work together to run through the test questions for each filter Digital Futures Manager ran through the test questions for each filter too
The Results Accuracy = the success rate of two things: 1) blocking what the filter is set to block 2) not blocking what the filter is set to allow through How good is the filter at doing its job?
Accuracy Across All Filters Average Filter Accuracy (margin of error +/- 5%)‏
Average Accuracy Ratings (Content) Average Filter Accuracy (margin of error +/- 5%)‏
Filtering Accuracy Summary Text-based web pages - 81% accuracy Images - 38% Email attachments - 25% RSS feeds - 53% Catalog searches - 67% Database searches - 83%
Overblocking on the Web WebMD the American Urological Association site VictimsOfPornography.org Univision.com DirtyPicturesBand.com Amazon and Google Book Search item pages TheSmokingGun.com Lesbian.org (a gay/lesbian support site)‏
More Overblocking on the Web the Wikipedia entry for Hustler Magazine a World War II history web site a UK breast cancer information site National Geographic images of beavers entire blogs are blocked because one of the many posts discussed something “adult” search results pages for a search for “Parents and Friends of Lesbians and Gays”
Overblocking in Library Resources a search for “orgasm” in the Health and Wellness Resource Center database a search for “vagina” in the World Book Encyclopedia online Searches for the following terms in our catalog: lesbianism, how to build a pipe bomb, sexual positions
Underblocking Getting through to images by clicking on the “full size” links Some adult search terms consistently allowed through (women’s asses, big penises)‏ Images of an adult sexual nature displayed on the search results page (some blocked, others not)‏
Workarounds Portal sites Clicking on “cached” version of webpage or image in search results Clicking on “full size” image link, even when thumbnail is blocked out Pages with images of an adult sexual nature but non-sexual text Plural word forms Pr0n instead of porn
Other findings Inconsistencies in the filtering of different text and image search engines The filtering programs do not handle non-English language words well Inconsistency in what’s blocked: “parents and lesbians” is blocked while “parents and gays” is allowed
Other Filtering Studies Review of studies from 2001-2008 Average accuracy of all studies of text-filtering = 78.56% Our accuracy average for text = 76.29% One image study accuracy = 48% Our accuracy average for imgages = 44%
We know what we're talking about!
What would we do differently? Test more non-English content Test video (popular sites, embedded players)‏ Test audio (popular sites, embedded players)‏ Test social sites (Facebook, LinkedIn)‏ Test Twitter Test pages using Ajax & other dynamic technologies
Conclusions All filters block a wide range of constitutionally protected content in an attempt to block other content.
More Conclusions Filters falsely block many valuable web pages and other online resources, on subjects ranging from war and genocide to safer sex and public health. No filter is reliably able to distinguish text or image content including obscenity, child pornography, or “harmful to minors” material from other, legal content.
What can we do with filters? We can block all images of all types on all web sites We can filter by keyword and URL in many categories, including categories with varying references to sexual content of an adult nature, etc., realizing there will be overblocking and underblocking
What can't we do with filters? We can not filter only images that are classified as obscene and harmful to minors.
How do I combat a proposal for filters? Use statistics from other studies Use good research techniques Use stories from your experiences
What if I have to filter? What are you trying to block out? Find out how various filters work Review ERATE funds impact Review cost & staff impact Review user impact Review past studies Decide how active you'll be unblocking
Questions? Sarah Houghton-Jan web: www.LibrarianInBlack.net IM: LibrarianInBlack Skype: LibrairanInBlack Facebook: facebook.com/librarianinblack Twitter: twitter.com/TheLiB email: [email_address]

More Related Content

Trying Not to Filter: Internet Filtering Technologies in Libraries

  • 1. Trying Not to Filter: Internet Filtering Technologies Update Sarah Houghton-Jan Digital Futures Manager for San Jose Public Library author of LibrarianInBlack.net This presentation will be at LibrarianInBlack.net
  • 2. San Jose Public Library's Filtering Challenge 1 City Council member + the Values Advocacy Council (& their $)‏ Proposed a new Internet Use Policy that would require filtering on all public computers Purpose stated: “to block websites that contain child pornography or material that is obscene”
  • 3. What is child pornography? What is harmful and obscene? There is no legal definition of pornography
  • 4. Federal Code & Case Law Applicable Court Cases Reno v. ACLU (1998), Miller v. California (1973), Smith v. United States (1977), Pope v. Illinois (1987)‏ Applicable Federal Law 18 U.S.C. Section 1470; 47 U.S.C. section 223(d) –Communications Decency Act of 1996, as amended by the PROTECT Act of 2003; 47 U.S.C. section 231 –Child Online Protection Act of 1998
  • 5. California State Code State Penal Code 311 defines obscene matter as “matter, taken as a whole, that to the average person, applying contemporary statewide standards, appeals to the prurient interest, that, taken as a whole, depicts or describes sexual conduct in a patently offensive way, and that, taken as a whole, lacks serious literary, artistic, political, or scientific value.”
  • 6. California State Code State Penal Code 313 defines harmful matter as “matter, taken as a whole, which to the average person, applying contemporary statewide standards, appeals to the prurient interest, and is matter which, taken as a whole, depicts or describes in a patently offensive way sexual conduct and which, taken as a whole, lacks serious literary, artistic, political, or scientific value for minors.”
  • 7. “ No filter, however, actually limits its categories to obscene material and child pornography because the current definition of obscenity doesn’t work on the internet.” ~ Lori Bowen Ayre, “Filtering and Filter Software”
  • 8. The Library’s Plan Work with the community to determine their position Assess cost and impact of filter implementation Internal policy review Review of other libraries’ practices Testing filters to see if they have improved
  • 9. Working with the Community Academic Senate from SJSU wrote a formal resolution against filtering Youth Commission – against Library Commission – against Public Feedback – against (about 3/4)‏
  • 10. Assessing Cost & Staffing Impact The money we'd get from government paled in comparison to the cost of implementation & maintenance. ERATE isn't worth it!
  • 11. Internal Policy & Other Libraries Existing computer use policy Talking with local libraries & libraries mentioned specifically by the city council member
  • 13. How Filters Work with URL List of trigger webpages Searches for trigger words in a popular search engine to form list (the black list!)‏ Some companies spot-check, some don't
  • 14. How Filters Work with Content List of trigger words Manually generated lists of words, based on thesauruses, combined with other factors (banner ads, numbers of images and/or links, etc.). Automated search process looks for a formula of the above then lists pages. Small # of filters filter based on file type
  • 15. Local Control Some filters let you form a white list of “allowed but normally blocked” content (URLs or keywords)‏ Some filters let you add to the blacklist Some filters let you temporarily unblock something for a customer.
  • 16. Major Filters on the Market 602LAN SUITE Content Filter 8e6 Barracuda Bess SmartFilter ContentProtect Professional Suite CyberPatrol CyberSentinel CyberSetting CyberSitter
  • 17. DP Inspector Dan’s Guardian DynaComm i:filter eSoft Web ThreatPak FastTracker FilterGate FilterPak FortiGate Series Image-Filter
  • 18. INternet Filter IF-2K iPrism Netmop MaxProtect NetNanny NetSentron NetSpective Webfilter NetSweeper Network Guardian
  • 19. Poesia SafeEyes Squidgard SurfControl Web Filter SurfPass WebSense Website-Echo WiseChoice
  • 20. What is being filtered, exactly? Filtering software companies do not tell their customers, in detail, the types of things they block in each category. Customers cannot obtain a complete list of words or URLs that are being blocked
  • 21. Choosing Filters to Test Working with IT Test computers on a test network Reviewing filter product information Reviewing filters being used in libraries Reviewing filter research and literature
  • 22. Filters We Tested WebSense Enterprise CyberPatrol Filtergate Barracuda
  • 23. What We Tested 135 questions/scenarios general keyword searches direct URL access image searches email text and photo attachments RSS feed content access searches in the online library catalog and databases
  • 24. How We Tested 2 librarians assigned to work together to run through the test questions for each filter Digital Futures Manager ran through the test questions for each filter too
  • 25. The Results Accuracy = the success rate of two things: 1) blocking what the filter is set to block 2) not blocking what the filter is set to allow through How good is the filter at doing its job?
  • 26. Accuracy Across All Filters Average Filter Accuracy (margin of error +/- 5%)‏
  • 27. Average Accuracy Ratings (Content) Average Filter Accuracy (margin of error +/- 5%)‏
  • 28. Filtering Accuracy Summary Text-based web pages - 81% accuracy Images - 38% Email attachments - 25% RSS feeds - 53% Catalog searches - 67% Database searches - 83%
  • 29. Overblocking on the Web WebMD the American Urological Association site VictimsOfPornography.org Univision.com DirtyPicturesBand.com Amazon and Google Book Search item pages TheSmokingGun.com Lesbian.org (a gay/lesbian support site)‏
  • 30. More Overblocking on the Web the Wikipedia entry for Hustler Magazine a World War II history web site a UK breast cancer information site National Geographic images of beavers entire blogs are blocked because one of the many posts discussed something “adult” search results pages for a search for “Parents and Friends of Lesbians and Gays”
  • 31. Overblocking in Library Resources a search for “orgasm” in the Health and Wellness Resource Center database a search for “vagina” in the World Book Encyclopedia online Searches for the following terms in our catalog: lesbianism, how to build a pipe bomb, sexual positions
  • 32. Underblocking Getting through to images by clicking on the “full size” links Some adult search terms consistently allowed through (women’s asses, big penises)‏ Images of an adult sexual nature displayed on the search results page (some blocked, others not)‏
  • 33. Workarounds Portal sites Clicking on “cached” version of webpage or image in search results Clicking on “full size” image link, even when thumbnail is blocked out Pages with images of an adult sexual nature but non-sexual text Plural word forms Pr0n instead of porn
  • 34. Other findings Inconsistencies in the filtering of different text and image search engines The filtering programs do not handle non-English language words well Inconsistency in what’s blocked: “parents and lesbians” is blocked while “parents and gays” is allowed
  • 35. Other Filtering Studies Review of studies from 2001-2008 Average accuracy of all studies of text-filtering = 78.56% Our accuracy average for text = 76.29% One image study accuracy = 48% Our accuracy average for imgages = 44%
  • 36. We know what we're talking about!
  • 37. What would we do differently? Test more non-English content Test video (popular sites, embedded players)‏ Test audio (popular sites, embedded players)‏ Test social sites (Facebook, LinkedIn)‏ Test Twitter Test pages using Ajax & other dynamic technologies
  • 38. Conclusions All filters block a wide range of constitutionally protected content in an attempt to block other content.
  • 39. More Conclusions Filters falsely block many valuable web pages and other online resources, on subjects ranging from war and genocide to safer sex and public health. No filter is reliably able to distinguish text or image content including obscenity, child pornography, or “harmful to minors” material from other, legal content.
  • 40. What can we do with filters? We can block all images of all types on all web sites We can filter by keyword and URL in many categories, including categories with varying references to sexual content of an adult nature, etc., realizing there will be overblocking and underblocking
  • 41. What can't we do with filters? We can not filter only images that are classified as obscene and harmful to minors.
  • 42. How do I combat a proposal for filters? Use statistics from other studies Use good research techniques Use stories from your experiences
  • 43. What if I have to filter? What are you trying to block out? Find out how various filters work Review ERATE funds impact Review cost & staff impact Review user impact Review past studies Decide how active you'll be unblocking
  • 44. Questions? Sarah Houghton-Jan web: www.LibrarianInBlack.net IM: LibrarianInBlack Skype: LibrairanInBlack Facebook: facebook.com/librarianinblack Twitter: twitter.com/TheLiB email: [email_address]